AI news ·
OpenAI Says Its AI Model Refuted the Exponential Time Hypothesis
OpenAI says an internal model disproved a major computer science theory, verified by a proof-checking tool, but independent experts are still reviewing the claim.

OpenAI announced that one of its internal models has generated a disproof of the exponential time hypothesis, a foundational conjecture in computer science. The company stated the proof was verified using the Lean proof assistant, a tool that checks mathematical arguments for correctness.
Independent researchers are now reviewing the claim. The result has not yet been confirmed by the wider scientific community, and OpenAI emphasized that extraordinary findings require external validation before they can be accepted.
The development highlights the potential for AI systems to contribute to advanced research while also demonstrating the continued need for rigorous human oversight. The review process is ongoing.